AI Paper Atlas

Research map / Adversarial robustness and security / Certified and empirical robustness

Batchnorm Allows Unsupervised Radial Attacks

Amur Ghose, Apurv Gupta, Yaoliang Yu, Pascal Poupart

NeurIPS 2023 · Adversarial robustness and security › Certified and empirical robustness · approach: Robust overfitting and regularization · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
angular attack vs FGSM, deeper network and larger epsAngular attack · ImageNet Resnet-152 eps=0.1top-1 accuracy: 9.82 -> 0.248Improved
angular attack vs FGSM on shallow net, small epsAngular attack · ImageNet Resnet-18 eps=0.03top-1 accuracy: 1.95 -> 23.416Mixed
angular attack vs PGD (supervised)Angular attack · ImageNet Resnet-18 eps=0.03top-1 accuracy: 0.006 -> 23.416Hurt
angular attack vs random noiseAngular attack · ImageNet Resnet-18 eps=0.03top-1 accuracy: 69.40 -> 23.416Improved
remove batchnorm (Fixup Resnet) instead of BN Resnet (ablation)Angular attack · ImageNet Fixup Resnettop-1 accuracyHurt
attack layer before BN layer instead of BN layer (ablation)Angular attack · ImageNet Resnet-18 eps=0.03top-1 accuracy: 23.416 -> 27.3Hurt
grant access to deeper blocks (8 to 12) (ablation)Angular attack · ImageNet Resnet-34 eps=0.03top-1 accuracy: 54.46 -> 6.40Improved
angular attack on ViT with LayerNorm vs FGSMAngular attack · ImageNet ViT-B-16 eps=0.1top-1 accuracy: 35.1 -> 16.7Improved
angular attack vs best single-stage attackAngular attack · ImageNet Resnet-18 eps=0.03top-1 accuracy: 30.45 -> 23.416Improved
vary initial iterations tinit 5-30 (ablation)Angular attack · ImageNet Resnets eps=0.03top-1 accuracyNo clear effect
increase total iterations from 30 to 50 (ablation)Angular attack · ImageNet Resnet-18 eps=0.03top-1 accuracy: -0.52 -> 0.38Improved
average angular loss over 4 layers instead of 1 (ablation)Angular attack · ImageNet Resnet-18 eps=0.03top-1 accuracy: -1.37 -> 3.49Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy